3 Replies Latest reply: May 15, 2014 10:03 AM by Grant Perkins RSS

    Open database question

    Chickenman _

      When opening a .csv, it throws the the error "First row contains bad data. This may indicate you need to check 'First row contains column names' option."

       

      And of course that's the case. The files more often than not contain hundreds of thousands or millions of rows so it sits a while and then moves to the sample data form, then sits a while after the box is checked.

       

      Seems to me this used to work automatically previously, but in any case is there a setting somewhere that will check the box rather than hanging up and making this a manual operation?

       

      CM

        • Open database question
          Grant Perkins

          If you open and select a Model at the same time presumably you will sail through most of the rest (relatively speaking for a large file ...)

           

          If the file is not previously modeled than I would be tempted to reduce it to a few limited set of representative records if you can and then develop an initial model with that. Once that looks OK load the full file and check for anomalies. If you can define the model with the assistance of the specification for the incoming file (field sizes and types for example) that would be even better.

           

          Or does this approach not cut the initial load effort these days?

           

           

          HTH.

           

           

          Grant

            • Open database question
              Chickenman _

              If you open and select a Model at the same time presumably you will sail through most of the rest (relatively speaking for a large file ...)

               

              If the file is not previously modeled than I would be tempted to reduce it to a few limited set of representative records if you can and then develop an initial model with that. Once that looks OK load the full file and check for anomalies. If you can define the model with the assistance of the specification for the incoming file (field sizes and types for example) that would be even better.

               

              Or does this approach not cut the initial load effort these days?

               

               

              HTH.

               

               

               

              Grant[/quote]

               

              Thank you for the reply, Grant. I'll take that as a "No" for getting the checkbox automatically checked. These tend to be ad hoc /Idumps from a large ERP database so no model would be practical.

               

              It's just mildly annoying to have to wait on the load, though it really isn't a very long wait. This in light of the fact that .xls load fine because the box IS checked.

               

              Best regards,

              CM

                • Open database question
                  Grant Perkins

                  I must confess I have not needed to consider this for database opening for a while but searching around I can't see anything that clearly indicates a default can be set. Still checking though.

                   

                  I'll guess that the Excel import is set based on the Excel workksheet parameters but even if not the worksheet will be a better defined data set than a txt based file can be so there is more for the interpreter to work with.

                   

                  I seem to recall that earlier versions only checked the first 250 lines but that threw up problems in some instances so generic performance improvements allowed the move to checking the entire report (I think). That's nothing much to do with the first line of course!

                   

                  Edit: Not too sure about the change in number of lines checked now Icome to consider it again. It may have been discussed but not implemented. V10 will report the total number of rows in the data source.

                   

                   

                  Grant